INDEX

Explanations

proper nouns, specifically names of authors and titles of their works

oai_token-act-pair · gpt-4o-mini Triggered by @bot

names of authors

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 kasarigan

-0.91

principalColumn

-0.69

ніципалі

-0.64

tinyos

-0.59

 ब्रेकडाउन

-0.54

 nahilalakip

-0.53

évaluateur

-0.51

uxxxx

-0.50

Diweddarwch

-0.50

 للمعارف

-0.48

POSITIVE LOGITS

Tracce

0.40

arParams

0.36

 useContext

0.35

ster

0.35

turi

0.34

 instead

0.34

SystemColors

0.33

 Gale

0.33

createClass

0.33

 mAdapter

0.32

Activations Density 0.052%

proper nouns, specifically names of authors and titles of their works

names of authors

No Comments

No Known Activations